v2.0a4

[2.0a4] - 2024-12-20

Fixed

• Fixed error parsing SelectedEventFunctionParams from JSON.
• Fixed error merging log space Histogram1D which resulted in infinite loops.

v2.0a3

[2.0a3] - 2024-12-11

Added

• Added SamplingParams.json_sampler_names, SamplingParams.append_to_sampler_names, SamplingParams.remove_from_sampler_names, SamplingParams.extend_sampler_names, SamplingParams.append_to_json_sampler_names, SamplingParams.remove_from_json_sampler_names, and SamplingParams.extend_json_sampler_names.
• Added jsonSampler and jsonSamplerMap to libcasm.monte.sampling.
• Added selected event data sampling methods to libcasm.monte.sampling.
• Added memory usage to results.

Changed

• Changed SamplingFixture::initialize to copy sampling functions from the SamplingFixtureParams object so functions that sample changes do not need to check if a new run has begun.

Fixed

• Fixed Conversions constructor to differentiate sublattices by symmetry and order of occupants. Documentation has been updated to reflect that the asym unit index used by Conversions, Mol, OccTransform, OccCandidate, OccCandidateList, etc. indicates sites that are equivalent by symmetry and order of occupants.

v2.0a2

[2.0a2] - 2024-07-17

Added

• Added to_json for CompletionCheckParams, SamplingFixtureParams, SamplingParams, jsonResultsIO
• Added "json_quantities" option to SamplingParams
• Added Conversions::species_list()

Changed

• Use shared_ptr to hold sampling fixtures in RunManager
• Output scalar quantities under "value" key in JSON results output
• Allow MethodLog to output to stdout
• Allow constructing libcasm.monte.ValueMap from dict

v2.0a1

[v2.0a1] - 2024-03-15

The libcasm-monte package provides useful building blocks for Monte Carlo simulations. This includes:

• Sampling classes and functions
• Equilibration, convergence checking, and statistics calculation
• Generic results IO
• Supercell index conversions
• Generic event definitions, construction, and selection

This package includes the Python package libcasm.monte, which may be installed via pip install, using scikit-build, CMake, and pybind11. This release also includes documentation, built using Sphinx.